Homework 1 (Advanced) Data Mining: Algorithms and Applications

Homework 1
(Advanced) Data Mining: Algorithms and Applications

 Datasets for this homework can be found at the link below and feel free to add this folder to your Google Drive:
https://drive.google.com/drive/u/0/folders/1ehWwunuAo7CE1Vk2JYkUnQMmxh5pph3C
1. Use ”Su_raw_matrix.txt” for the following questions (30 points).
(a) Use read.delim function to read Su_raw_matrix.txt into a variable called su. (Notice that su has become a
data frame now)
(b) Use mean and sd functions to find mean and standard deviation of Liver_2.CEL column.
(c) Use colMeans and colSums functions to get the average and total values of each column.
2. Use rnorm(n, mean = 0, sd = 1) function in R to generate 10000 numbers for the following (mean, sigma) pairs
and plot histogram for each, meaning you need to change the function parameter accordingly. Then comment on how
these histograms are different from each other and state the reason. (20 points)
(a) mean=0, sigma=0.2
(b) mean=0, sigma=0.5
Please save your figures as image from RStudio. (Hint: to see the difference in plots you may need to set the xlim
parameter in plot function to c(-5,5))
3. Perform the steps below with ”dat” dataframe which is just a sample data for you to observe how each plot function
( 3b through 3e ) works. Notice that you need to have ggplot2 library installed on your system. Please refer slides how
to install and import a library. Installation is done only once, but you need to import the library every time you need it
by saying library(ggplot2). Then run the following commands for questions from 3a through 3e and observe how
the plots are generated first. (40 points)
(a) dat <- data.frame(cond = factor(rep(c("A","B"), each=200)),
rating = c(rnorm(200),rnorm(200, mean=.8)))
(b) # Overlaid histograms
ggplot(dat, aes(x=rating, fill=cond)) +
geom_histogram(binwidth=.5, alpha=.5, position="identity")
(c) # Interleaved histograms
ggplot(dat, aes(x=rating, fill=cond)) + geom_histogram(binwidth=.5, position="dodge")
(d) # Density plots
ggplot(dat, aes(x=rating, colour=cond)) + geom_density()
(e) # Density plots with semitransparent fill
ggplot(dat, aes(x=rating, fill=cond)) + geom_density(alpha=.3)
(f) Read ”diabetes_train.csv” into a variable called diabetes and apply the same functions 3b through 3e for the
mass attribute of diabetes and save the images. (Hint: instead of cond above, use the class attribute to color
your groups. When you have fill option, your plots should show same type of chart for both groups in different
colors on the same figure. Keep in mind that diabetes and dat are both DataFrames)
4. By using quantile(), calculate 10th
, 30th
, 50th
, 60th percentiles of skin attribute of diabetes data. (10 points)
 Important
• Please put all images and their explanations in a single pdf file and submit it along with an R script which has all R
commands that you used.
• Please leave comments in your R script
• Only submissions through Canvas will be accepted.
• If you have a GitHub page and want to submit an R notebook, you can give a link to your work on GitHub
1